Fault location with DGs in radial distribution system using radial basis function neural network

Abstract

Increasing penetration of Renewable Energy in energy market will contribute to increasing number of distributed generation (DG) existence in the grid, thus leading to conflict in fault location, detection and protection coordination in distribution system. This study is focuses on single-line-to-ground (LG) fault detection in a radial distribution system. The objective of this study is to estimate fault location in a radial distribution system in the presence of DGs by using Radial Basis Function Neural Network (RBFNN), with consideration to minimize monitor placement in system. Fault location has been estimated in term of faulty bus. Two types of radial distribution network with DGs have been tested in this study; 10 bus and 34 bus network. Fault analysis has been performed using Power World simulator and data generated has been applied for RBFNN development via MATLAB. RBFNN performance was then evaluated statistically, by SSE, R2 and RMSE. The proposed RBFNN has been able to accurately predict current magnitude at unmonitored buses by only few provided monitored buses readings. With accurate predicted results by the neural network, pattern of current magnitude during fault has been observed in order to identify faulty buses. It was shown in this study that faulty bus can be identified 100% using the proposed approach.